1. Field of the Invention
This invention relates generally to data processing systems, and more particularly to data processing systems in which groups of characters, called data strings, are transferred between storage units and a central processing unit. In certain data formats, one character of a group of characters can have a sign designating that a character is superimposed, or overpunched, on a decimal character. Because the overpunched format can not be utilized by an execution unit in the central processing unit, it is necessary to alter the character with the overpunched data entering the central processing unit. The disclosed invention identifies an overpunched character in an incoming character group, converts the overpunched character to a normal character by removing the superimposed signals designating a sign character, and inserts the normal character into the correct position in the incoming data. The portion of the central processing unit responsible for the execution of instruction is thereby able to utilize the character.
2. Description of the Prior Art
The storing of a numerical quantity in a memory subsystem generally requires that the sign of the quantity be retained in a selected character of the data string. One method of retaining the sign of the quantity in a data string is to superimpose, or overpunch, the signals representing the quantity sign on the signals representing a decimal value in a selected data character position. However, in order to achieve a format which is usable by an execution unit of a central processing unit, it is necessary to "correct" the overpunched character by separating the decimal signals from the sign signals and by reinserting the number or corrected signals in the character position which was formerly overpunched. Otherwise, the manipulation of the data string in the execution will generally result in a meaningless operand.
It was known in the prior art to determine this position of the character of the data string containing the sign, i.e., an overpunched character digit, and then correct the signed character by means of a series of firmware-controlled steps. Generally, the firmware-controlled steps require an unacceptable length of time and compromise the performance of a data processing system.
It is therefore an object of the present invention to provide an improved data processing system.
It is a further object of the present invention to "correct" an incoming data string so that the resulting data string can be utilized in the execution of decimal operations.
It is a still further object of the present invention to provide a decimal unit in a data processing system which converts a decimal data string with sign information in one character position to a decimal data string without the sign information.
it is yet another object of the present invention to determine the position of the character to be replaced in a data word and to replace the character with a character capable of being utilized in decimal operations.
It is still another object of the present invention to replace (with a multiplicity of microprogrammed instructions) a character, having sign information, in an incoming data string with a character representing the absolute value of the signal character.